MARINE MAMMAL SCIENCE, 20(3):527-545 (July 2004) 0 2004 by the Society for Marine Mammalogy DIETS OF COASTAL BOTTLENOSE DOLPHINS FROM THE U.S. MID-ATLANTIC COAST DIFFER BY HABITAT DAMONP. GANNON' DANIELLEM. WAPLES Duke University Marine Laboratory, Nicholas School of the Environment and Earth Sciences, Duke University, 135 Duke Marine Lab Road, Beaufort, North Carolina 28516, U.S.A. E-mail: [email protected] ABSTRACT We recorded 31 species in the stomachs of 146 coastal bottlenose dolphins (Tursiops truncatu) from North Carolina, U.S.A. Sciaenid fishes were the most common prey (frequency of occurrence = 95%). By mass, Atlantic croaker (Micropogonias undulatus) dominated the diet of dolphins that stranded inside estuaries, whereas weakfish (Cynosicon regalis) was most important for dolphins in the ocean. Inshore squid (Loligo sp.) was eaten commonly by dolphins in the ocean, but not by those in the estuaries. There was no significant pattern in prey size associated with dolphin demography, but the proportion of the diet represented by croaker was higher for males than for females, and mature dolphins ate more croaker than did juveniles. Dietary differences between dolphins that stranded in the estuaries and those that stranded on ocean beaches support the hypothesis that some members of the population inhabit the ocean primarily while others reside principally in estuaries. The overwhelming majority of prey were soniferous species (75% of numerical abundance), which is consistent with the hypothesis that bottlenose dolphins use passive listening to locate noise-making fishes. However, spatiotemporal patterns in consumption of Sciaenid fishes did not coincide with their spawning, which is when peak sound production is thought to occur. Key words: coastal bottlenose dolphin, Tursiops trzncatus, food habits, feeding, foraging ecology, Sciaenidae, soniferous fishes, mid-Atlantic, estuary, coastal ocean. In this paper we quantify the food habits of coastal bottlenose dolphins, Tursiups truncatus, in the waters off the mid-Atlantic coast of the U.S. using traditional stomach content analysis techniques. Previous analyses of bottlenose dolphin food habits in this region were based on small samples, and none investigated differences in feeding habits associated with season, habitat, sex, or age. True (1885) identified common gurnard (Prionutus carolinus) in the stomachs of two dolphins taken by the dolphin fishery in Cape May, New Jersey. Fishermen from Cape Hatteras, North Carolina, told Townsend (1914) that the bottlenose dolphins in that region Current address: Mote Marine Laboratory, 1600 Ken Thompson Parkway, Sarasota, FL 34236, U.S.A. 527 5 28 MARINE MAMMAL SCIENCE, VOL. 20, NO. 3, 2004 primarily ate seatrout (Cynoscion spp.). Mead and Potter (1990) conducted the most thorough investigation of bottlenose dolphin food habits in the mid-Atlantic region, using a sample of 64 dolphins stranded on the coasts of North Carolina, Virginia, and Maryland. These authors calculated frequencies of prey occurrence and numerical abundance, both of which require large sample sizes, and neither accounts for prey mass (Hyslop 1980). The most important prey found by Mead and Potter (1990) included spot (Leiostomus xantburzls), croaker (Micropogonias undulatus), seatrouts, silver perch (Bairdiella cbtysoura), hake (Uropbycis sp.), and squid (Loligo sp.). We investigated variation in food habits with respect to habitat (coastal ocean us. estuary), season, and demography. In North Carolina bottlenose dolphins move between the ocean and the estuaries.* However, there is evidence from photographic identification and the presence of ectoparasites that some dolphins occupy estuaries primarily, while others stay mainly in the ocean. Some authors hypothesize that pseudostalked barnacles Xenobalanus sp., which attach to the appendages of dolphins, can be used as an indicator of ocean residence (Caldwell and Caldwell 1972, Barros 1993). If some coastal dolphins are primarily oceanic while others are primarily estuarine, one might expect to find differences between the two groups with regard to food habits. We tested the hypothesis that the dolphin diet changes seasonally because many of the species eaten by bottlenose dolphins undergo seasonal migrations. Faced with migrating prey, dolphins can either follow their prey or switch to a different set of prey. Developing the skills to locate and capture prey probably requires an extended period of learning for young bottlenose dolphins. Mann and Smuts (1999) noted that young dolphins begin “play snacking” behavior as early as three weeks of age with objects such as seagrass. Play snacking with fish begins at about week seven. While developing hunting skills, young dolphins may eat prey that are less elusive than those species normally eaten by older dolphins. Smith and Read (1992) found that harbor porpoise (Pbocoena pbocoena) calves eat euphausiids (MeganyctiFbanes norvegica), which are rare in the diets of adult porpoises. Young franciscana (Pontoporia blainvillei) feed on shrimp, whereas adults eat fishes and cephalopods (Rodriguez et al. 2002). Mead and Potter (1990) also concluded that weaning is gradual in bottlenose dolphins, which may reflect the need for young dolphins to augment energy intake from solid food with that of their mother’s milk until they have become proficient at capturing prey. Lactating female mammals typically have the highest energy demands of any demographic group within their species and often have nutrient requirements that differ from other members of their species (e.g., Clutton-Brock et al. 1982, Close et al. 1985, Perez and Mooney 1986). In a comparison of the diets of pregnant and lactating spotted dolphins (Stenella attenuata), Bernard and Hohn (1989) found that lactating spotted dolphins ate more food and increased the proportion of fish relative to squid in their diet. Robertson and Chivers (1997) also found that female spotted dolphins increased their food consumption during lactation. However, Robertson and Chivers (1997) came to the opposite conclusion regarding the relative propor- tions of squid and fish in the diets of lactating females: lactating females increased the amount of squid in their diet. Cockcroft and Ross (1990) found dietary dif- ferences among life history stages of bottlenose dolphins in South Africa. Recently, Kastelein et al. (2002) documented that captive bottlenose dolphins increased their ’ Personal communication from Keith Rittmaster, North Carolina Maratime Museum, 3 15 Front Street., Beaufort, NC 28516, June 2003. GANNON AND WAPLES: BOTTLENOSE DOLPHIN DIET 529 consumption of food little during gestation but that their consumption increased by 58%-97% during lactation. METHODS We examined the stomach contents of coastal bottlenose dolphins that died of natural causes or that were captured accidentally by commercial fisheries in the estuaries and ocean waters of North Carolina from 1993 to 2001. Two ecotypes of ?: truncatus are recognized on the east coast of the U.S.: a nearshore form known as the “coastal” ecotype and a pelagic form known as the “offshore”ecotype (see Hersh and Duffield 1990; Mead and Potter 1990, 1995; Torres et al. 2003). We restricted our analyses to the coastal form. Offshore dolphins are not thought to come any closer to shore than 7.5 km and coastal dolphins are found from the estuaries out to a distance of 34 km from shore (Torres et al. 2003). We used a hierarchical approach to distinguish between members of the two ecotypes. Dolphin stomachs containing prey known to inhabit only waters greater than 25 m were excluded from analysis (Mead and Potter 1995). The remaining dolphins were categorized on the basis of parasite assemblages. Coastal dolphins were identified by the presence of the trematode Braunina in pyloric and main stomach chambers and offshore dolphins were recognized by the cestodes Pbyllobotbrium and Monorbygma and by the nematode Crassicauda (Mead and Potter 1990, 1995). Categorization of some dolphins was corroborated independently by sighting histories and capture locations. Photo- graphic sighting data of individual dolphins (n = 10) were collected by the contributors to the Mid-Atlantic Bottlenose Dolphin Cata10g.~These ten dolphins were only sighted within 7.5 km of shore and were, therefore, assumed to be of the coastal ecotype. The location of capture was known for seven dolphins; five were captured in pelagic fisheries (>34 km from shore: offshore ecotype) and two were taken in coastal fisheries (c7.5 km from shore: coastal ecotype). The stomach of each dolphin was extracted in its entirety (including esophagus and duodenal ampulla) during necropsies conducted either in the field or in the laboratory. In the lab, the contents of all three stomach chambers were removed for analysis. Intact prey were removed first, then the remaining stomach contents were put in a shallow tray. We topped the tray off with water, gently agitated it, and then decanted floating tissue into a 1-mm sieve. We repeated this process until all floating tissue was separated from the dense bones. We retained the skeletal parts that had sunken to the bottom of the tray and then inspected the contents of the sieve to make sure no hard parts, such as small otoliths, had adhered to the loose flesh that had been poured off. Structures used to identify prey included: sagittal otoliths, dentary bones, and skulls of teleosts; lower mandibles from cephalopods; and exoskeletons from crustaceans. Prey items were identified with